基于视觉的非完整约束移动机器人位姿控制研究

发布时间：2018-04-28 05:26

本文选题：非完整约束 + 移动机器人　；参考：《哈尔滨工业大学》2017年硕士论文

【摘要】：随着移动机器人技术日趋成熟,移动机器人在工业生产中扮演越来越重要的角色。尤其在物流、仓库管理等领域,轮式移动机器人具有巨大的应用前景。为了缩短开发周期及成本,轮式移动机器人通常由舵轮驱动的电动叉车改装而来,这类移动机器人的控制问题比较复杂。一方面机器人的轮子只能滚动不能沿轴向滑动,移动机器人的运动受到约束,属于非完整约束系统,导致移动机器人的位置、航向控制问题比较复杂。另一方面改装之后,舵角控制系统对输入指令跟踪精度不足,导致机器人控制精度不足。本文设计了移动机器人的位姿控制算法,并给出了控制器参数的选取原则。通过在极坐标系下建立移动机器人的运动学模型,设计了线性定常的状态反馈控制律,实现机器人的位置误差、航向误差同时收敛。通过构造李雅普诺夫函数对系统的局部稳定性进行了证明,给出了系统稳定范围。通过对控制器参数进一步优化,为了增加系统的稳定范围,当移动机器人的初始条件不满足局部稳定条件时,机器人会先运动到局部稳定区域内,保证系统最终可以渐进稳定。通过仿真验证了算法的有效性。针对移动机器人舵角控制精度不足,本文提出了舵角控制系统的复合校正方案,提升了舵角控制系统对输入指令的跟踪精度。由电动叉车改装而来的机器人舵角控制系统的跟踪精度不高,输出存在严重滞后,如果不经校正,移动机器人的控制误差不能满足使用需求。为了提升舵角控制系统对输入指令的跟踪精度,采用按输入补偿的复合校正结构对舵角控制系统进行校正。为了合理选取控制器参数,对舵角控制系统进行了建模,通过扫频法得到舵角控制系统的幅频特性曲线和相频特性曲线,并对模型参数进行了辨识。由于按输入补偿的复合校正方法中需要引入前馈信号,论文根据移动机器人位姿控制器解析式,给出了舵角指令导数,得到前馈环节所需的输入信号导数。由于不需要对指令差值除以时间的方式求微分,在提高了舵角控制系统的跟踪精度的同时也保证系统的稳定性。经过输入补偿复合校正后,移动机器人的位置控制误差小于8 mm,航向控制误差小于0.17 rad。
[Abstract]:With the development of mobile robot technology, mobile robot plays a more and more important role in industrial production. Especially in the fields of logistics and warehouse management, wheeled mobile robots have great application prospects. In order to shorten the development cycle and cost, wheeled mobile robots are usually refitted from electric forklifts driven by steering wheels. The control problems of this kind of mobile robots are complex. On the one hand, the wheel of the robot can only roll and can not slide along the axis, the movement of the mobile robot is constrained and belongs to the nonholonomic constraint system, which leads to the position of the mobile robot and the course control problem is more complicated. On the other hand, the rudder angle control system has insufficient tracking accuracy for input instructions, which leads to the lack of robot control accuracy. In this paper, the position and attitude control algorithm of mobile robot is designed, and the selection principle of controller parameters is given. By establishing the kinematics model of mobile robot in polar coordinate system, a linear constant state feedback control law is designed to realize the robot's position error and heading error converging simultaneously. The local stability of the system is proved by constructing Lyapunov function, and the stability range of the system is given. Through further optimization of controller parameters, in order to increase the stable range of the system, when the initial condition of the mobile robot does not satisfy the local stability condition, the robot will first move into the local stable region to ensure that the system can eventually be asymptotically stable. The effectiveness of the algorithm is verified by simulation. Aiming at the insufficient precision of rudder angle control of mobile robot, a compound correction scheme of rudder angle control system is proposed in this paper, which improves the tracking accuracy of rudder angle control system to input instructions. The tracking accuracy of the rudder angle control system of the robot modified from the electric forklift is not high, and the output is seriously delayed. If the control error of the mobile robot is not corrected, the control error of the mobile robot can not meet the requirements. In order to improve the tracking accuracy of rudder angle control system, the rudder angle control system is corrected by a compound correction structure based on input compensation. In order to select the controller parameters reasonably, the rudder angle control system was modeled. The amplitude-frequency characteristic curve and phase frequency characteristic curve of rudder angle control system were obtained by sweep frequency method, and the model parameters were identified. Due to the need to introduce feedforward signal into the compound correction method of input compensation, according to the analytical formula of position and attitude controller of mobile robot, the rudder angle instruction derivative is given, and the input signal derivative of feedforward link is obtained. Since it is not necessary to divide the differential value by time, the tracking accuracy of the rudder angle control system is improved and the stability of the system is guaranteed. After the compound correction of input compensation, the position control error of the mobile robot is less than 8 mm and the heading control error is less than 0.17 rad.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP242

【参考文献】